Amiga.org

Amiga computer related discussion => Amiga Software Issues and Discussion => Topic started by: alphonsus on October 24, 2002, 12:16:50 AM

Title: Benchmarking WinUAE
Post by: alphonsus on October 24, 2002, 12:16:50 AM
I've just upgraded to an Athlon XP 1800+ (it's quite fast).
In testing my emulation to see how fast it is, I've discovered sysinfo is broken when trying to display too high a MHz rating for a chip. AIAB seems to be broken as well on the emulation, despite people posting modules for it. Any idea where I can get a benchmark prog that works so I can feel smug?
(I haven't bought an AmigaONE yet as waiting for a large cash injection - though I'm not sure entirely why I'm apologising for WinUAE)
Incidentally, benchmarking a 500MHz celeron gave me a 264Mhz 020 - quite an efficient emulation - even if the test is biased.
Title: Re: Benchmarking WinUAE
Post by: kd7ota on October 24, 2002, 12:18:55 AM
For some friggin reason, its REALLY slow on this 500mhz AMD system.  I never could get WinUAE to go fast on this.  I dont know what program you can use to see the mhz on WinUAE...... Anyone else know? :-)
Title: Re: Benchmarking WinUAE
Post by: Piru on October 24, 2002, 12:31:53 AM
MHz reported under the emulation have nothing to do with real performance.
Do some real stuff, like heavy gfx or video processing and time that on both real Amiga and emulation.
It will give you more realistic idea of the true speed of th emulation.
Title: Re: Benchmarking WinUAE
Post by: bhoggett on October 24, 2002, 12:40:27 AM
That kind of comparison is useless. The only way to get a reasonable idea of how fast your emulation is running is to do "real-life" benchmarks. In other words, work out a number of tasks using various applications, then time them with a stopwatch.

Benchmark programs give bogus results on UAE because the timing is all wrong and inconsistent. Even on Amithlon, the results won't be accurate because the benchmark programs expect Classic Amiga hardware and performance and go a bit crazy when the results fall outside those parameters.
Title: Re: Benchmarking WinUAE
Post by: alphonsus on October 24, 2002, 09:03:01 PM
I _know_ that it won't yeild a real-life score - I did say I wanted to do it for smug value.
It does yeild information in one way - comparing the emulation with _itself_! I use sysinfo to test what effects any changes I make to the settings have on the emulation.
Title: Re: Benchmarking WinUAE
Post by: on October 24, 2002, 09:09:14 PM
Hmm I was pritty sure AIAB did include one that worked, if not 2. I'd have to look to see what they were. Perhaps you're just too darn fast :-D
Title: Re: Benchmarking WinUAE
Post by: Piru on October 24, 2002, 09:12:18 PM
Ok, use SysSpeed 2.6 (http://www.aminet.net/util/moni/sspeed26.lha)
Title: Re: Benchmarking WinUAE
Post by: alphonsus on November 01, 2002, 11:32:19 AM
Thanks for the info - I'm trying sysspeed this evening
Title: Re: Benchmarking WinUAE
Post by: Quixote on November 01, 2002, 11:41:56 AM
;-) Piru and BHogget are correct to point out that the most useful benchmark for UAE is to run software on it.  

:-( For example, Imagine will not run on my installation of MacUAE, probably because of something that isn't implimented, such as the floating point unit or something.  Since Imagine falls down on MacUAE, it doesn't matter how fast the emulated processor is.